Spatial multi-criteria frameworks for urban physical resilience against earthquakes combine expert judgment with statistical adjustment to weight indicators of urban structure, but their weights have never been validated against real post-earthquake outcomes. The source framework (Meimandi Parizi, Taleai and Sharifi, 2024) structures twenty urban-structure indicators with Interpretive Structural Modeling, weights them by the Analytic Network Process, and adjusts those expert weights with bootstrap confidence intervals, Shannon entropy, and constrained optimization, applied to Local Climate Zone units of Tehran District 4. Because no damaging earthquake has struck the study area, the authors could only run sensitivity analysis and scenario-making as internal checks, leaving open whether the weights — or weights from alternative derivations such as discrete-choice models — actually predict observed seismic damage, service disruption, and recovery. The open problem is to establish an externally validated weighting scheme for this framework: confront the resilience classifications induced by its own expert-based weights, or by any competing derivation, with independent empirical post-earthquake data, and determine which weights, if any, yield classifications predictive of observed damage, service disruption, and recovery.

Background

Urban physical resilience denotes the capacity of a city's built fabric — buildings, street networks, open spaces — to withstand a hazard, maintain function, and recover. Assessment frameworks operationalize this by scoring spatial units on a set of indicators across the phases of disaster risk management (preparedness, absorption, recovery, adaptation). A widespread approach is multi-criteria decision analysis: the source study (Meimandi Parizi, Taleai and Sharifi, Natural Hazards 120:13041–13076, 2024) first uses Interpretive Structural Modeling (ISM), an expert-judgment procedure that organizes pairwise influence relations among indicators into a hierarchical structure, and then the Analytic Network Process (ANP), a generalization of the Analytic Hierarchy Process (AHP, the standard technique for deriving weights from expert pairwise comparisons among criteria arranged in a strict hierarchy) that permits interdependence among criteria and derives indicator weights from expert pairwise-comparison matrices. Because purely expert-based weights are vulnerable to judgment bias, the framework adjusts them with objective statistical and optimization techniques: bootstrapping (resampling expert inputs to obtain confidence intervals for each weight), Shannon-entropy weighting (assigning objective weights from the information content of the indicator values across zones), and a constrained optimization problem whose solution is combined multiplicatively with the ANP weights. The spatial unit is the Local Climate Zone (LCZ), a classification of urban terrain into types defined by building density, height, and land cover — originally developed for urban-climate studies and reused here as a physically homogeneous analysis unit. Applied to twenty indicators in District 4 of Tehran, the framework yields phase-wise resilience scores for each LCZ zone, with sensitivity analysis and scenario-making as internal robustness checks. Two limitations define the open question. First, since no damaging earthquake has occurred in the study area, the scores could not be compared with any real post-earthquake damage, service-disruption, or recovery data, so the external validity of the weights is unknown. Second, the authors note that alternative weight-adjustment methods — they name discretized choice models, i.e. discrete-choice models, statistical models of choice behavior estimated from stated or revealed preference data — could yield more accurate weights; but 'more accurate' is only meaningful relative to observed outcomes, which loops back to the missing empirical validation. The same authors' earlier papers (Integrated methods, Nat. Hazards 2021; GIS-based multi-criteria framework, Sustainability 2022) developed the ISM structuring and the bootstrap–entropy–optimization adjustment without external validation either.

Problem Statement

Establish an externally validated weighting scheme for the phase-based spatial framework for urban physical resilience of Meimandi Parizi, Taleai and Sharifi (2024). The target is whether indicator weights for this twenty-indicator, ISM-structured framework — its own expert-based ISM/ANP weights (with or without its bootstrap–entropy–optimization adjustment), or any competing derivation — yield phase-wise resilience classifications predictive of observed earthquake consequences. The weights and the phase-wise resilience classifications they induce must be confronted with independent empirical post-earthquake outcome data (observed building damage, service disruption, and/or recovery trajectories) at a comparable spatial resolution; since Tehran District 4 has no such data, the framework is transferred to an earthquake-prone city for which outcome data from a documented post-earthquake event are available, and locating and selecting that city and dataset is itself part of the task. The determination is which weighting scheme, if any, yields resilience scores that best predict the observed outcomes. The objective is to replace internally sensitivity-checked expert weights with weights whose predictive validity against real earthquake consequences is established; because external validity is the object of investigation, a determination that no weighting scheme's weights predict the observed outcomes is itself a valid resolution of the open problem.

References

  1. Meimandi Parizi S, Taleai M, Sharifi A (2024) A spatial evaluation framework of urban physical resilience considering different phases of disaster risk management. Natural Hazards 120(14):13041–13076. DOI 10.1007/s11069-024-06703-0. https://doi.org/10.1007/s11069-024-06703-0
  2. Parizi SM, Taleai M, Sharifi A (2022) A GIS-Based Multi-Criteria Analysis Framework to Evaluate Urban Physical Resilience against Earthquakes. Sustainability 14(9):5034. DOI 10.3390/su14095034. https://doi.org/10.3390/su14095034
  3. Parizi SM, Taleai M, Sharifi A (2021) Integrated methods to determine urban physical resilience characteristics and their interactions. Natural Hazards 109(1):725–754. DOI 10.1007/s11069-021-04855-x. https://doi.org/10.1007/s11069-021-04855-x
  4. Lalith Prakash E, Anand KB, Jena R, Pradhan B (2026) A holistic approach to seismic risk by integrating preparedness into hybrid deep learning and ANP-based spatial modeling: A case study of Coimbatore city, India. Sustainable Cities and Society 149:107699. DOI 10.1016/j.scs.2026.107699. https://doi.org/10.1016/j.scs.2026.107699
  5. Chen Y, Xie W, Xu Y, Wang D, Dong Q (2026) Uncovering the Foundation-Capacity-Performance logic of urban resilience: identifying dynamic weights and nonlinear relationships through a spatial machine learning framework. Sustainable Cities and Society 140:107239. DOI 10.1016/j.scs.2026.107239. https://doi.org/10.1016/j.scs.2026.107239
  6. Alikaei S, Bakhtiar Nasrabadi A (2025) Towards a Resilient Urban Future: Implementing Risk-Sensitive Physical Urban Planning in Tehran's Earthquake-Prone District 10 Based on Scenario Building Technique. International Journal of Environmental Research 19(6):236. DOI 10.1007/s41742-025-00873-w. https://doi.org/10.1007/s41742-025-00873-w
  7. Jiang D, Li Z (2026) Design of Urban resilience model in emergency situations. PLoS ONE 21(6):e0347588. DOI 10.1371/journal.pone.0347588. https://doi.org/10.1371/journal.pone.0347588
  8. Ha Dinh Nghiem, Sharifi A, Aghaloo K, Haque MN (2026) Indicators for Assessing Urban Resilience to Climate Change. Annual Review of Environment and Resources. DOI 10.1146/annurev-environ-121124-015717. https://doi.org/10.1146/annurev-environ-121124-015717